Pith. sign in

REVIEW 3 major objections 5 minor 43 references

Multimodal Visual Image Based User Association and Beamforming Using Graph Neural Networks

T0 review · 3 major / 5 minor · reviewed 2026-08-15 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read Pairing camera views with RF pilots lets a graph neural network assign users and design beams within about 7.6% of the perfect-CSI exhaustive-search minimum rate.

desk verdict A serious vision-plus-pilot GNN system for joint association and beamforming, but the headline numbers rest on a signal-model equation that uses the transpose instead of the Hermitian inner product, so the results need a mandatory check before they can be trusted. read the letter →

arxiv 2506.18218 v2 pith:NF56UYF6 submitted 2025-06-23 eess.SP

Yinghan Li , Yiming Liu , Wei Yu This is my paper · ORCID
classification eess.SP
keywords multimodallearninguserassociationbeamforminggraphneuralnetworkmax-minfairnessvisualdataRFpilotsunsupervised
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

The paper claims that a wireless cellular network can solve the joint user-association and beamforming problem under max-min fairness without estimating explicit channel state information. Instead, a detection network extracts user locations from camera images, two graph neural networks independently process those locations and a small number of RF pilots, and a third multimodal GNN fuses the two feature streams to output association decisions and beamforming vectors. On a simulated urban intersection with ray-traced channels, the multimodal GNN (GNN-C) consistently beats both the pilot-only and image-only GNNs and comes within about 7.6% of the minimum rate achieved by exhaustive search with perfect CSI. The appeal is that visual data provides channel awareness at no spectral cost, so pilots can be kept short.

What carries the argument

The load-bearing object is a three-stage multimodal graph neural network. Each user is a node in a fully connected graph; GNN-I takes estimated locations from images, GNN-P takes correlated received pilots $\bar{r}_{kl}$, and GNN-C concatenates the refined node features from both and outputs candidate associations and beamforming vectors. The normalization block maps outputs to feasible sets, and a modified straight-through estimator (STE) with an annealing coefficient $\lambda$ treats discrete association decisions as differentiable while preventing user–AP pairs from being dropped from gradient updates too early. A detection network with filtering and camera calibration converts pixels to world coordinates, and the whole pipeline is trained unsupervised by minimizing the negative minimum rate plus a penalty that prevents premature association lock-in.

What would settle it

A concrete test: deploy the trained detector in a real intersection without color-coded vehicles, collect synchronized images and pilots, and compare GNN-C's downlink minimum rate against GNN-P's. If the multimodal gain disappears or reverses once location errors exceed about one meter or detections are missed, the central claim would be refuted.

Watch

Extended reading notes

Core claim

The central claim is that multimodal integration of visual location estimates and RF pilots lets a learning-based system approach the performance of perfect-CSI exhaustive search for the joint user association and downlink beamforming problem, under a max-min fairness objective. The paper argues that the channel can be written as $h = M(\hat{Q}) + \hat{N}$, a deterministic function of user locations plus an independent residual, so locations derived from images carry substantial channel information. The proposed GNN-C fuses features separately extracted by an image-based GNN and a pilot-based GNN, and is trained unsupervised with a modified straight-through estimator that keeps beamforming vectors trainable while user associations gradually harden into binary decisions. In simulations with $K=6$ users and $L=4$ access points, GNN-C achieves a downlink minimum rate within about 7.6% of exhaustive search with perfect CSI, outperforming pilot-only and image-only models as well as a direct multimodal GNN baseline.

Load-bearing premise

The whole approach rests on the premise that a channel can be adequately reconstructed from user locations plus an independent residual, and that a mapping learned in a simulated, color-coded intersection will transfer to real urban deployments where detection is less reliable.

Editorial extensions

If this is right

  • With camera coverage, the system can operate with far fewer pilots than traditional CSI-based optimization, since location features compensate for corrupt or missing channel estimates.
  • The same GNN structure generalizes across different numbers of users and pilot lengths; training with 64 pilots nearly matches the full standard model for inference pilot counts from 16 to 1024.
  • A single network instance runs in milliseconds (about 0.17 ms for feature processing plus 4.6 ms for detection), compared with minutes for exhaustive search, making real-time association and beamforming possible.
  • The learned user associations often follow load balancing rather than nearest-AP assignment, indicating that the network internalizes fairness-aware decisions.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • A natural next step is to test whether the multimodal gain survives when users are not color-coded or when cameras are partially occluded; if location errors grow beyond the roughly one-metre accuracy seen in simulation, the image branch may add noise instead of information.
  • The separate feature-extraction design suggests a modular deployment: operators who already have pilot-based GNNs could add an image branch without retraining the pilot branch from scratch.
  • The same framework could be extended to include AP-side permutation equivariance, or to fuse other sensing modalities such as radar or LiDAR, which provide location information under lighting conditions where cameras fail.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

3 major / 5 minor

Summary. The paper proposes a learning-based method for joint user association and beamforming in a downlink MU-MISO system under a max-min fairness criterion, using multimodal data: camera images and uplink RF pilots. A detection network (YOLOv5 plus camera calibration) extracts user locations from images; two separate GNNs, GNN-I and GNN-P, process image-derived locations and pilots, respectively; a third GNN, GNN-C, fuses their features to output binary user associations and beamforming vectors. Training is unsupervised with respect to optimal labels, using the max-min rate as the loss and a modified straight-through estimator to handle the binary association constraint. Simulations in a CARLA/Sionna environment with four four-antenna APs and six car-like UEs show that GNN-C outperforms pilot-only and image-only GNNs, approaches the performance of exhaustive search with perfect CSI (about a 7.6% rate reduction), and generalizes across numbers of users and pilot lengths. The paper also reports low inference complexity and provides a visual interpretation of beamforming and association decisions.

Significance. If the reported results hold, the paper makes a useful contribution to vision-aided wireless optimization: it demonstrates that multimodal location-plus-pilot inputs can drive a GNN toward a near-optimal max-min-fair association and beamforming solution without explicit CSI estimation, with a strong baseline set including exhaustive search, LMMSE-based methods, and fusion ablations. The paper deserves credit for providing a code link, using realistic ray-traced channels, and including generalization and interpretability studies. However, the numerical claims are contingent on correcting the complex inner-product convention in the rate expression, and the real-world relevance is weakened by the color-coded-car detection assumption. The central idea is plausible and the simulation methodology is generally sound, but the load-bearing rate expression and the detection assumptions need to be addressed before the results can be fully appraised.

major comments (3)
  1. [II.A, Eqs. (4)-(5); VI, Eq. (28)] The received signal and rate expressions use h_{kl}^T v_{kl} and |h_{kl}^T v_{kl}|^2. For complex channels, the physical received signal is h_{kl}^H v_{kl}, and |h_{kl}^T v_{kl}|^2 is not the achievable signal power. Since Sionna generates complex channel coefficients, the objective in (6) and all rates reported in Fig. 6 are not the true max-min rates unless the code computes a different expression from the one in the paper. The interpretation section later uses p(θ,φ)^H v in Eq. (28), which is internally inconsistent with the channel inner product convention. This issue is load-bearing for the headline claim of a 7.6% gap to perfect-CSI exhaustive search: if the simulation follows the equations, the comparison is against a fictitious objective; if the code uses the Hermitian inner product, the equations should be corrected so the results can be appraised. Please correct the convention and state clearly that all reported rates are computed with h^H v.
  2. [III.A and V.A] The detection stage assumes that UEs are cars with unique colors and class labels ('each car is assigned a distinct class label' in Section III.A; 'each car is assigned a unique color for identification' in Section V.A). This assumption removes the user-identification and data-association problem that is central in real vision-based systems, and the CARLA simulator does not model occlusion, lighting variation, or detection confusions that would break color-based identification. The paper claims suitability for practical deployment in Section VII without stating this limitation. Please either add an explicit limitations paragraph and temper the deployment claim, or evaluate robustness when users are not uniquely colored and when detectors miss or confuse users.
  3. [IV.A.5 and V.F] The training loss is described as using 'the estimated rate R_hat_k^I from (5)', but the paper never states which channel h is used to compute the rate inside the loss. In the simulation one can use the true Sionna channel; at inference only pilots and images are available. Since the paper's contribution is to avoid explicit CSI estimation, please specify explicitly: (i) whether the training loss for GNN-P and GNN-C is evaluated with true channel realizations or with pilot-derived channel estimates, and (ii) how the evaluation rates in Fig. 6 are computed (true h, estimated h, or a learned channel representation). This is needed to reproduce the results and to confirm that the comparisons against perfect-CSI and LMMSE baselines are fair.
minor comments (5)
  1. [Section I.C] The sentence defining the notation (·)^I, (·)^P, (·)^C is repeated verbatim; one copy should be deleted.
  2. [Section IV.B] The input feature vector for GNN-P is not defined in the text; from Fig. 3 and Table II it appears to be the concatenation of the real and imaginary parts of the correlated pilot vectors, but this should be stated explicitly.
  3. [Section V.C] The test set contains only 200 samples; the paper should report confidence intervals or error bars for the minimum-rate results in Figs. 6 and 7, since the reported differences (especially the 7.6% gap) may be sensitive to this sample size.
  4. [Section V.G, Table III] The running-time table lists detection time for GNN-I and GNN-C but not for GNN-P; please clarify whether the reported pilot-based GNN time excludes image detection, and make the table entries consistent in format.
  5. [Section IV and IV.C] There are typos: 'archiecture' in the first paragraph of Section IV and 'associaiton' in Section IV.C; these should be corrected.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity: the GNN is trained directly on the max-min fairness objective and benchmarked against external exhaustive search with perfect CSI, so the central claim is self-contained.

full rationale

The paper's derivation chain is self-contained. The system model in Eqs. (4)-(5) defines the max-min rate objective, the image detection network maps pixels to locations (Section III), and the GNNs map locations and pilots to association and beamforming decisions (Section IV). Training is unsupervised with a loss that is the negative of the objective plus a penalty term, and the reported performance is measured on a separate test set (Section V.F) against exhaustive search with perfect CSI, LMMSE estimation, and strongest-channel assignment. No fitted parameter is later renamed as a prediction: the hyperparameters lambda and lambda_p are training stabilizers with scheduled schedules, not parameters tuned to the reported 7.6% gap. The only self-citation, reference [21] by co-author Wei Yu, appears in the related-work discussion about pricing-based user association and is not load-bearing for any derivation or benchmark. The location-to-channel assumption h = M(Q_hat) + N_hat in Eq. (11) is explicitly stated as a modeling assumption rather than being derived from the target result. The transpose-vs-Hermitian issue raised in the skeptic's take is a potential correctness concern about the rate metric, not a circularity: all compared methods use the same metric, so it does not make any prediction equivalent to its input by construction. Therefore the paper does not exhibit self-definitional, fitted-input, or self-citation-driven circularity.

Assumptions & free parameters 4 free parameters · 5 assumptions · 0 invented entities

The central claim rests on the learnability of a location-to-channel mapping from simulated data and on the fidelity of the CARLA/Sionna pipeline. The free parameters are training hyperparameters and calibration assumptions; no new physical entities are introduced.

free parameters (4)
  • STE balancing coefficient lambda = not specified (scheduled, increasing over training)
    Controls the trade-off between continuous and binary user association in Eq. (22); the schedule is not given.
  • penalty coefficient lambda_p = not specified (decreasing over time)
    Weight on the user-association penalty term in the loss (Section IV.A.5); no schedule specified.
  • detection confidence threshold = 0.7
    Filtering block in Section III rejects detections below 0.7 confidence; chosen by hand.
  • assumed car center height z_hat_w = not specified (fixed antenna height 0.4 m, car center height estimated)
    Used in camera calibration Eq. (16); error in this height directly biases location estimates.
assumptions (5)
  • domain assumption Block-fading channel with TDD reciprocity, so uplink pilots give downlink CSI.
    Section II.A; the whole pilot-based branch relies on this.
  • domain assumption The channel is a deterministic function of UE locations plus an independent residual, h = M(Q_hat) + N_hat.
    Equation (11) in Section II.B; the multimodal gain rests on this.
  • domain assumption Ray-tracing simulation (Sionna) and CARLA renderings are faithful proxies for real propagation and detection.
    Sections V.A and V.B; all ground-truth channels and images come from these simulators.
  • ad hoc to paper UEs are uniquely identifiable by distinct car colors and class labels.
    Section III.A; enables YOLO detection but is unrealistic for general users.
  • domain assumption Max-min fairness is the appropriate objective.
    Standard in the related works cited; the paper optimizes this and benchmarks against it.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Multimodal Visual Image Based User Association and Beamforming Using Graph Neural Networks." pith.science (2026). https://pith.science/paper/NF56UYF6

@misc{pith2026250618218,
  author       = {Pith},
  title        = {Pith review of: Multimodal Visual Image Based User Association and Beamforming Using Graph Neural Networks},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/NF56UYF6}},
  note         = {Machine review of arXiv:2506.18218}
}
read the original abstract

This paper proposes an approach that leverages multimodal data by integrating visual images with radio frequency (RF) pilots to optimize user association and beamforming in a downlink wireless cellular network under a max-min fairness criterion. Traditional methods typically optimize wireless system parameters based on channel state information (CSI). However, obtaining accurate CSI requires extensive pilot transmissions, which lead to increased overhead and latency. Moreover, the optimization of user association and beamforming is a discrete and non-convex optimization problem, which is challenging to solve analytically. In this paper, we propose to incorporate visual camera data in addition to the RF pilots to perform the joint optimization of user association and beamforming. The visual image data helps enhance channel awareness, thereby reducing the dependency on extensive pilot transmissions for system optimization. We employ a learning-based approach based on using first a detection neural network that estimates user locations from images, and subsequently two graph neural networks (GNNs) that extract features for system optimization based on the location information and the received pilots, respectively. Then, a multimodal GNN is constructed to integrate the features for the joint optimization user association and beamforming. Simulation results demonstrate that the proposed method achieves superior performance, while having low computational complexity and being interpretable and generalizable, making it an effective solution as compared to traditional methods based only on RF pilots.

Figures

Figures reproduced from arXiv: 2506.18218 by the authors.

Figure 1
Figure 1. Architecture of the image detection network. [PITH_FULL_IMAGE:figures/full_fig_p005_1.png] view at source ↗
Figure 2
Figure 2. Three coordinate systems used in the camera calibrat [PITH_FULL_IMAGE:figures/full_fig_p006_2.png] view at source ↗
Figure 3
Figure 3. Architecture of the proposed GNN network. [PITH_FULL_IMAGE:figures/full_fig_p008_3.png] view at source ↗
Figures from the paper (5 more)
Figure 4
Figure 4. Figure 4: (a) The 3D model of the simulation environment in the C [PITH_FULL_IMAGE:figures/full_fig_p010_4.png]
Figure 5
Figure 5. Figure 5: Errors of estimated locations from the image detecti [PITH_FULL_IMAGE:figures/full_fig_p011_5.png]
Figure 6
Figure 6. Figure 6: Minimum rate performance of the proposed GNN with [PITH_FULL_IMAGE:figures/full_fig_p012_6.png]
Figure 7
Figure 7. Figure 7: Generalization Performance of the Proposed GNN Mode [PITH_FULL_IMAGE:figures/full_fig_p013_7.png]
Figure 8
Figure 8. Figure 8: Array response of the APs and user association decisi [PITH_FULL_IMAGE:figures/full_fig_p014_8.png]

Review history (2 revisions) →

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

43 extracted references · 40 canonical work pages

  1. [27]

    Computer vision aided beam t racking in a real-world millimeter wave deployment,

    S. Jiang and A. Alkhateeb, “Computer vision aided beam t racking in a real-world millimeter wave deployment,” in Proc. IEEE Globecom W orkshops (GC Wkshps), Dec. 2022, pp. 142–147

    work page 2022

  2. [31]

    Multi-user ma tching and resource allocation in vision aided communications,

    W. Xu, F. Gao, Y . Zhang, C. Pan, and G. Liu, “Multi-user ma tching and resource allocation in vision aided communications,” IEEE Trans. Commun., vol. 71, no. 8, pp. 4528–4543, Aug. 2023

    work page 2023

  3. [1]

    A survey on resou rce allocation for 5G heterogeneous networks: Current researc h, future trends, and challenges,

    Y . Xu, G. Gui, H. Gacanin, and F. Adachi, “A survey on resou rce allocation for 5G heterogeneous networks: Current researc h, future trends, and challenges,” IEEE Commun. Surveys Tuts. , vol. 23, no. 2, pp. 668–695, 2nd Quart. 2021

    work page 2021

  4. [2]

    Channel estim ation techniques based on pilot arrangement in OFDM systems,

    S. Coleri, M. Ergen, A. Puri, and A. Bahai, “Channel estim ation techniques based on pilot arrangement in OFDM systems,” IEEE Trans. Broadcast., vol. 48, no. 3, pp. 223–229, Sep. 2002

    work page 2002

  5. [3]

    Inductive repres entation learning on large graphs,

    W. Hamilton, Z. Ying, and J. Leskovec, “Inductive repres entation learning on large graphs,” Adv. Neural Inf. Process. Syst. (NeurIPS) , vol. 30, pp. 1024–1034, 2017

    work page 2017

  6. [4]

    User association in 5G networks: A survey and an outlook,

    D. Liu et al. , “User association in 5G networks: A survey and an outlook,” IEEE Commun. Surv. Tutor . , vol. 18, no. 2, pp. 1018–1044, 2nd Quart. 2016

    work page 2016

  7. [5]

    Goldsmith, Wireless Communications

    A. Goldsmith, Wireless Communications. Cambridge, U.K.: Cambridge Univ. Press, 2005

    2005

  8. [6]

    Energy- efficient user association in cognitive heterogeneous netw orks,

    A. Mesodiakaki, F. Adelantado, L. Alonso, and C. V erikou kis, “Energy- efficient user association in cognitive heterogeneous netw orks,” IEEE Commun. Mag. , vol. 52, no. 7, pp. 22–29, Jul. 2014

    work page 2014

Show all 43 references
  1. [7]

    Dynamic cell as sociation for downlink sum rate maximization in multi-cell heterogeneou s networks,

    S. Corroy, L. Falconetti, and R. Mathar, “Dynamic cell as sociation for downlink sum rate maximization in multi-cell heterogeneou s networks,” in Proc. IEEE Int. Conf. Commun. (ICC) , 2012, pp. 2457–2461

    2012

  2. [8]

    Joint downlink base statio n as- sociation and power control for max-min fairness: Computat ion and complexity,

    R. Sun, M. Hong, and Z. Q. Luo, “Joint downlink base statio n as- sociation and power control for max-min fairness: Computat ion and complexity,” IEEE J. Sel. Areas Commun. , vol. 33, no. 6, pp. 1040– 1054, Jun. 2015

    2015

  3. [9]

    Joint base stat ion association and power control via benders’ decomposition,

    L. Qian, Y . J. A. Zhang, Y . Wu, and J. Chen, “Joint base stat ion association and power control via benders’ decomposition, ” IEEE Trans. Wireless Commun., vol. 12, no. 4, pp. 1651–1665, Apr. 2013

    2013

  4. [10]

    Joint uplink base st ation asso- ciation and power control for small-cell networks with non- orthogonal multiple access,

    L. Qian, Y . Wu, H. Zhou, and X. Shen, “Joint uplink base st ation asso- ciation and power control for small-cell networks with non- orthogonal multiple access,” IEEE Trans. Wireless Commun. , vol. 16, no. 9, pp. 5567–5582, Sep. 2017

    2017

  5. [11]

    User associa tion with maximizing weighted sum energy efficiency for massive MIMO- enabled heterogeneous cellular networks,

    T. Zhou, Z. Liu, D. Qin, N. Jiang, and C. Li, “User associa tion with maximizing weighted sum energy efficiency for massive MIMO- enabled heterogeneous cellular networks,” IEEE Commun. Lett. , vol. 21, no. 10, pp. 2250–2253, Oct. 2017

    2017

  6. [12]

    Integrated power control and base station assignment,

    R. Y ates and C.-Y . Huang, “Integrated power control and base station assignment,” IEEE Trans. V eh. Technol. , vol. 44, no. 3, pp. 638–644, Aug. 1995

    1995

  7. [13]

    Downl ink power control and base station assignment,

    F. Rashid-Farrokhi, K. R. Liu, and L. Tassiulas, “Downl ink power control and base station assignment,” IEEE Commun. Lett. , vol. 1, no. 4, pp. 102–104, Jul. 1997

    1997

  8. [14]

    Optimal join t base station assignment and beamforming for heterogeneous networks,

    M. Sanjabi, M. Razaviyayn, and Z.-Q. Luo, “Optimal join t base station assignment and beamforming for heterogeneous networks,” IEEE Trans. Signal Process. , vol. 62, no. 8, pp. 1950–1961, Apr. 2014

    1950

  9. [15]

    Joint user association and resource allocation in small cell networks with backhaul constraints,

    Z. Cui and R. Adve, “Joint user association and resource allocation in small cell networks with backhaul constraints,” in Proc. Annu. Conf. Inf. Sci. Syst. (CISS) , Mar. 2014, pp. 1–6

    2014

  10. [16]

    De ep reinforce- ment learning for user association and resource allocation in heteroge- neous networks,

    N. Zhao, Y .-C. Liang, D. Niyato, Y . Pei, and Y . Jiang, “De ep reinforce- ment learning for user association and resource allocation in heteroge- neous networks,” in Proc. IEEE Global Commun. Conf. (GLOBECOM) , Dec. 2018, pp. 1–6

    2018

  11. [17]

    Energy efficient user association and power allocation in m illimeter- wave-based ultra dense networks with energy harvesting bas e stations,

    H. Zhang, S. Huang, C. Jiang, K. Long, V . C. M. Leung, and H . V . Poor, “Energy efficient user association and power allocation in m illimeter- wave-based ultra dense networks with energy harvesting bas e stations,” IEEE J. Sel. Areas Commun. , vol. 35, no. 9, pp. 1936–1947, ...

    1936

  12. [18]

    Distributed energy efficient fair user association in mass ive MIMO enabled HetNets,

    D. Liu, L. Wang, Y . Chen, T. Zhang, K. K. Chai, and M. Elkas hlan, “Distributed energy efficient fair user association in mass ive MIMO enabled HetNets,” IEEE Commun. Lett. , vol. 19, no. 10, pp. 1770–1773, Oct. 2015

    2015

  13. [19]

    α-fairness-maximizing user association in energy-constrained small cell networks,

    J. Jang and H. J. Y ang, “ α-fairness-maximizing user association in energy-constrained small cell networks,” IEEE Trans. Wireless Com- mun., vol. 21, no. 9, pp. 7443–7459, Sep. 2022

    2022

  14. [20]

    Joint user association and eICIC for max–min fairness in HetNets,

    Y . Jia, M. Zhao, and W. Zhou, “Joint user association and eICIC for max–min fairness in HetNets,” IEEE Commun. Lett. , vol. 20, no. 3, pp. 546–549, Mar. 2016

    2016

  15. [21]

    Distributed pricing-based user asso ciation for downlink heterogeneous cellular networks,

    K. Shen and W. Y u, “Distributed pricing-based user asso ciation for downlink heterogeneous cellular networks,” IEEE J. Sel. Areas Com- mun., vol. 32, no. 6, pp. 1100–1113, Jun. 2014

    2014

  16. [22]

    Joint user associatio n and hybrid beamforming designs for cell-free mmWave MIMO communicati ons,

    Z. Wang, M. Li, R. Liu, and Q. Liu, “Joint user associatio n and hybrid beamforming designs for cell-free mmWave MIMO communicati ons,” IEEE Trans. Commun. , vol. 70, no. 11, pp. 7307–7321, Nov 2022

    2022

  17. [23]

    Max-mi n SINR coordinated multipoint downlink transmission—duality an d algorithms,

    D. W. H. Cai, T. Q. S. Quek, C. W. Tan, and S. H. Low, “Max-mi n SINR coordinated multipoint downlink transmission—duality an d algorithms,” IEEE Trans. Signal Process. , vol. 60, no. 10, pp. 5384–5395, Oct. 2012

    2012

  18. [24]

    Applying deep-lear ning-based com- puter vision to wireless communications: Methodologies, o pportunities, and challenges,

    Y . Tian, G. Pan, and M.-S. Alouini, “Applying deep-lear ning-based com- puter vision to wireless communications: Methodologies, o pportunities, and challenges,” IEEE Open J. Commun. Soc. , vol. 2, pp. 132–143, 2021

    2021

  19. [25]

    Deep le arning based channel covariance matrix estimation with user locat ion and scene images,

    W. Xu, F. Gao, J. Zhang, X. Tao, and A. Alkhateeb, “Deep le arning based channel covariance matrix estimation with user locat ion and scene images,” IEEE Trans. Commun. , vol. 69, no. 12, pp. 8145–8158, Dec. 2021

    2021

  20. [26]

    Vision-aid ed 6G wireless communications: Blockage prediction and proactive handof f,

    G. Charan, M. Alrabeiah, and A. Alkhateeb, “Vision-aid ed 6G wireless communications: Blockage prediction and proactive handof f,” IEEE Trans. V eh. Technol., vol. 70, no. 10, pp. 10 193–10 208, Oct. 2021

    2021

  21. [28]

    Comput er vision aided mmwave beam alignment in V2X communications,

    W. Xu, F. Gao, X. Tao, J. Zhang, and A. Alkhateeb, “Comput er vision aided mmwave beam alignment in V2X communications,” IEEE Trans. Wireless Commun., vol. 22, no. 4, pp. 2699–2714, Apr. 2023

    2023

  22. [29]

    Multi-camer a views based beam searching and bs selection with reduced training overhead,

    B. Lin, F. Gao, Y . Zhang, C. Pan, and G. Liu, “Multi-camer a views based beam searching and bs selection with reduced training overhead,” IEEE Trans. Wireless Commun. , vol. 72, no. 5, pp. 2793–2805, May 2024

    2024

  23. [30]

    Vision-position multi-modal beam prediction using real m illimeter wave datasets,

    G. Charan, T. Osman, A. Hredzak, N. Thawdar, and A. Alkha teeb, “Vision-position multi-modal beam prediction using real m illimeter wave datasets,” in Proc. IEEE Wireless Commun. Netw. Conf. (WCNC) , Apr. 2022, pp. 2727–2731

    2022

  24. [32]

    Interference management and user as sociation for nested array-based massive mimo hetnets,

    M. Feng and S. Mao, “Interference management and user as sociation for nested array-based massive mimo hetnets,” IEEE Trans. V eh. Technol., vol. 67, no. 1, pp. 454–466, Jan. 2018

    2018

  25. [33]

    Advanced user associa tion in non-orthogonal multiple access-based fog radio access net works,

    L. Qi, M. Peng, Y . Liu, and S. Y an, “Advanced user associa tion in non-orthogonal multiple access-based fog radio access net works,” IEEE Trans. Commun., vol. 67, no. 12, pp. 8408–8421, Dec. 2019

    2019

  26. [34]

    Joint power al location and user association optimization for irs-assisted mmwave systems,

    D. Zhao, H. Lu, Y . Wang, H. Sun, and Y . Gui, “Joint power al location and user association optimization for irs-assisted mmwave systems,” IEEE Trans. Wireless Commun. , vol. 21, no. 1, pp. 577–590, Jan. 2022

    2022

  27. [35]

    Comparative analysis of deep learning im age detection algorithms,

    S. Srivastava, A. V . Divekar, C. Anilkumar, I. Naik, V . K ulkarni, and V . Pattabiraman, “Comparative analysis of deep learning im age detection algorithms,” J. Big Data , vol. 8, no. 1, p. 66, 2021

    2021

  28. [36]

    ultralytics/yolov5: v7.0 - YOLOv5 SOTA Realtime Instance Segmentation,

    Ultralytics, “ultralytics/yolov5: v7.0 - YOLOv5 SOTA Realtime Instance Segmentation,” https://github.com/ultralytics/yolov5.com, 2022. [Online]. Available: https://doi.org/10.5281/zenodo.7347926

    2022 doi

  29. [37]

    Camera calibration with distortion models and accuracy evaluation,

    J. Weng, P . Cohen, M. Herniou et al., “Camera calibration with distortion models and accuracy evaluation,” IEEE Trans. Pattern Anal. Mach. Intell., vol. 14, no. 10, pp. 965–980, Oct. 1992

    1992

  30. [38]

    ENGNN: A general edg e- update empowered GNN architecture for radio resource manag ement in wireless networks,

    Y . Wang, Y . Li, Q. Shi, and Y .-C. Wu, “ENGNN: A general edg e- update empowered GNN architecture for radio resource manag ement in wireless networks,” IEEE Trans. Wireless Commun. , vol. 23, no. 6, pp. 5330–5344, Jun. 2024

    2024

  31. [39]

    Learning power allocation for multi -cell-multi- user systems with heterogeneous graph neural networks,

    J. Guo and C. Y ang, “Learning power allocation for multi -cell-multi- user systems with heterogeneous graph neural networks,” IEEE Trans. Wireless Commun., vol. 21, no. 2, pp. 884–897, Feb. 2022

    2022

  32. [40]

    Carla: An open urban driving simulator,

    A. Dosovitskiy, G. Ros, F. Codevilla, A. Lopez, and V . Ko ltun, “Carla: An open urban driving simulator,” in Proc. Conf. Robot Learn. , 2017, pp. 1–16

    2017

  33. [41]

    Sionna: An open-source library for next- generation physical layer research,

    J. Hoydis et al. , “Sionna: An open-source library for next- generation physical layer research,” 2023. [Online]. Avai lable: https://arxiv.org/abs/2203.11854

    2023 arXiv

  34. [42]

    GNN-based powe r allocation and user association in digital twin network for the terahertz 16 band,

    H. Zhang, X. Ma, X. Liu, L. Li, and K. Sun, “GNN-based powe r allocation and user association in digital twin network for the terahertz 16 band,” IEEE J. Sel. Areas Commun. , vol. 41, no. 10, pp. 3111–3121, Oct. 2023. Yinghan Li (Graduate Student Member, IEEE) re- ceived the B...

    2023

  35. [2023]

    Wei Yu (Fellow, IEEE) received the B.A.Sc

    He was also recognized as an Exemplary Reviewer for IEE E WIRELESS COMMUNICATIONS LETTERS in 2022. Wei Yu (Fellow, IEEE) received the B.A.Sc. degree in computer engineering and mathematics from the University of Waterloo, Waterloo, ON, Canada, and the M.S. and Ph.D. degrees in...

    2022

Pith tools

Reviewed August 15, 2026 · model on record in the stance chip above.